Diagnose indexing issues with bulk URL Inspection

The scenario

Half your new blog posts aren't showing up on Google. Search Console's URL Inspection works on one URL at a time, manually. With 200+ posts in a content cluster, that's 200 clicks, 200 wait-for-results, 200 copy-pastes. You want a one-shot bulk audit that tells you which specific URLs are crawled-not-indexed, which are discovered-not-indexed, and which fail Mobile Usability.

The prompt

Use the gsc-pap MCP. For https://example.com, first call get_index_coverage_summary to see the high-level distribution of indexing states. Then list the top 50 most-recently-published URLs in /blog/* (you can ask me to provide these or pull them from a sitemap). Run batch_inspect_urls on those URLs. Group results by `indexingResult.coverageState`: how many are INDEXED vs CRAWLED_NOT_INDEXED vs DISCOVERED_NOT_INDEXED vs other? For any not indexed, give me the URL + the specific verdict + the most likely fix.

What happens

1. 1

   AI gets the account-wide picture first

   `get_index_coverage_summary` returns counts of pages in each indexing state across all your verified properties. This sets the baseline, if you have 10k pages and 200 are unindexed, that's normal; if you have 500 and 200 are unindexed, that's a content-engineering problem.

2. 2

   AI runs batch_inspect_urls on the candidate set

   You provide the URL list (from a sitemap, recent publish list, or a content cluster). The tool batches up to 2,000 URLs in a single call. Results include lastCrawlTime, indexingState, mobile usability, structured-data validity, all per URL.

3. 3

   AI clusters and prescribes

   Output is the bucketed list with a one-line fix per pattern: discovered-not-indexed → likely thin content or low internal linking; crawled-not-indexed → likely a quality signal; mobile-usability fail → specific viewport / tap-target / font-size issue. The AI prioritises by URL importance (you can tell it 'these are revenue pages').

Outcome

Most content sites discover one repeating pattern (e.g. 80% of unindexed posts share the same low-link, thin-intro template). Fix the template, re-request indexing, watch the bulk inspect run a week later show the bucket emptying. What used to be a 'we'll get to it eventually' problem becomes a 30-minute weekend project.

Where to take it next

• ·Pair with `list_sitemaps` + `get_sitemap_details` to confirm the URLs are at least submitted via sitemap.
• ·After fixes ship, re-run batch_inspect_urls to confirm the indexing state moved.
• ·Build a Claude Code script that re-runs this audit every two weeks and emails you only the diff.